Mid-Pandemic Sentiment: A Story of Time and Place
Additional Comments Without Non-Responses
- Initially it was obvious most people don’t fill out this part of the survey, but filtering reveals there might be some general timeliness issues, or other need for better time partitioning.
What Improved the Experience?
- Again it looks like time is a big factor, lessons and planning show up here a lot too.
What Went Well?
- “Collaboration” appears to be working well, and “discussions” appears as well, group based interaction seems very successful. “Time” makes yet another appearance as well.
Most Common Two Word Sequences
- The most common bigram by far is “more time”, “was very”, and “was great” also show up, indicating people are doing a great job.
Sample of Responses with “more time”
- REPLACE WITH UPDATED INTERACTIVITY
A Network of the Data
- Model of most common bigrams based on eigenvector centrality: fancy math way of ranking importance of different connections which are denoted by “nodes” - the blue circles, and arrows to indicate “direction” of the relationship.
AFINN Model
AFINN gives a numeric sentiment value for each word, with positive or negative numbers to indicate sentiment.
- The sentiment is subsequently weighted by number of occurrences.
Word Topic Probability Model
Per-topic-per-word probabilities topic model with a number of topics proportional to the response types.
- Note again the appearance of more time as top responses to “what could have improved your experience?”